YC Summer 2026 (S26) founder

Aman Raj

Founder of Decawork, Training company-specific models that approve AI actions.

Company

Aman Raj is listed as a founder of Decawork. An agent’s access should depend on the task it is doing. The same action can be appropriate for one task and wrong for another. Access decisions need to adapt in real time as the agent works, but companies can’t turn every business judgment into a fixed rules or have people manually approve every step. We train small, company-specific models that approve AI actions in real time. They consider the company’s policies and interests, the agent’s current task, and its past actions to decide whether each proposed action should be allowed. Access is specific to the task at hand and reassessed as the agent works. Companies will rent intelligence from AI labs or run open-source models to power their agents. They’ll own the models that keep their entire AI workforce acting in their interests. Decawork is building those models.

Founder traction evidence

Public posts and activity attributed directly to this founder.

  1. LinkedIn

    At 22 I was convinced Barclays moved slowly because nobody there understood technology.

    At 22 I was convinced Barclays moved slowly because nobody there understood technology. It was my first job out of college. And it took me some time to see what was actually happening at that bank. Every tool had a person attached to it whose neck was on the line if it did...

  2. LinkedIn

    "The agent is fine.

    "The agent is fine. Tell me what it CAN'T do." An IT head said this to me barely 10 minutes into a demo. Cut me off mid-sentence. I didn't have a good answer that day. Since then I've done 30+ calls with IT and security leaders, and some version of that question comes up every...

  3. X

    How do you know you are building something that people want? @_sarthak4

    How do you know you are building something that people want? @_sarthak4

  4. LinkedIn

    A working demo doesn't mean your AI agent is ready to employ.

    A working demo doesn't mean your AI agent is ready to employ. It proves capability. Employment is a different bar. Companies are already deploying agents that handle messy workflows, touch internal data, and take real actions inside their systems. And the moment one has a goal,...

  5. LinkedIn

    "It's quite mind-blowing that all of this happened autonomously!"

    "It's quite mind-blowing that all of this happened autonomously!" That's the Hugging Face CEO, describing how his company got hacked by OpenAI's own AI. An agent had a narrow testing goal and, to hit it, used stolen credentials and a vulnerability nobody knew existed to break...

  6. X

    instantly doubled their revenue in last 2 years.

    instantly doubled their revenue in last 2 years. so yes, people are still replying to cold emails ig.

  7. LinkedIn

    Somewhere in your stack, an ex-employee still has a working login.

    Somewhere in your stack, an ex-employee still has a working login. You just haven't found the app yet. Their SSO is dead, and the big apps are clean because SCIM handled those. The miss lives in the long tail - apps that never supported SCIM, or lock it behind an enterprise...

  8. X

    orkut and songs[.]pk I feel like a boomer lol

    orkut and songs[.]pk I feel like a boomer lol